The 2026 Trustees Report put a date on it: the fourth quarter of 2032, six years out, at which point Social Security’s retirement trust fund is depleted, and every check is automatically cut by 22%. Not phased in. Not means-tested. An across-the-board cut of roughly $500 a month for the average beneficiary, hitting the 70 million people the program currently pays.
The standard framing is demographic destiny… too few workers per retiree, an aging country, arithmetic nobody can vote against. The demographics are real. But the framing hides two design choices sitting in plain sight, one at each end of the income distribution.
Wages above $184,500 are exempt from the tax that funds the program. Undocumented workers pay $26 billion a year into benefits they are barred from ever claiming. Nothing paid at the top; everything paid and nothing owed back at the bottom.
Start at the top. The Social Security payroll tax stops at the taxable maximum, which means the program taxes 100% of a nurse’s wages and a shrinking fraction of a hedge fund manager’s. Roughly $3.8 trillion in earnings escape the tax entirely, as much as $475 billion a year in forgone revenue by one analysis of the SSA’s own wage data, with institutional estimates like the Peterson Foundation’s clustering around $3 trillion per decade. As income has concentrated above the cap, the share of national wages subject to the tax has drifted from 87% down to about 83%. The exemption is not a law of nature. It is a line item.
Now the bottom. ITEP’s comprehensive study found undocumented immigrants paid $96.7 billion in total taxes in a single year, and $25.7 billion of it went specifically to Social Security, a program they cannot access.1 In actuarial terms, these are the program’s best customers: pure contribution, zero liability. And the trustees themselves cite reduced immigration as one of the reasons the depletion date moved earlier this year.
Mass deportation is not only a humanitarian question. It is a direct withdrawal from the trust fund.
So I built the ledger as a model you can operate.
Three levers: revenue from taxing above-cap wages, a 25% tax on equity-backed borrowing dedicated to the trust fund, and undocumented worker contributions on a slider that runs from mass deportation (−$26B/year) to work authorization (+$12B/year, because authorized workers earn more and comply more). The model is calibrated to the trustees’ trajectory; current law lands exactly on Q4 2032 and a 22% cut. Then you move the levers and watch the date move.
Full cap removal buys decades of runway, not permanence, surplus for several years, then structural deficits resume. The model shows that honestly, because the actuaries’ own finding is that uncapping closes about two-thirds of the long-run shortfall, not all of it.
The honest caveats are on the page. Taxing above-cap earnings without benefit credit breaks the contribution-to-benefit link the program was designed around; that is the strongest objection, and the model names it.
But that link is already broken in the other direction: the workers at the bottom of this ledger have no link at all. And the equity-loan lever breaks it deliberately; in the other direction, it is the smallest lever in dollars, but the only one that taxes capital rather than labor to fund a program built on labor.
Move all three levers, and the point becomes hard to unsee… the people being told the program cannot afford them are subsidizing it, and the people who could most easily fund it are exempt from doing so.
The 2032 cliff is a scheduled event with published solutions. What is missing is not arithmetic.
Sources: 2026 OASDI Trustees Report summary — OASI depletion Q4 2032, 78% of benefits payable thereafter; ITEP, Tax Payments by Undocumented Immigrants — $96.7B total, $25.7B to Social Security in 2022; Peter G. Peterson Foundation, on eliminating the tax cap; Congressional Research Service, RL32896. The model is deliberately simple — a baseline deficit calibrated to the trustees’ trajectory, adjusted by three levers. It is meant to make the structure of the choice operable, not to replace the actuaries.